On Sat, 30 Jul 2016 17:49:59 -0700, Jeff Liebermann wrote:
About the same range. At least the same range within some
reasonable tolerance range, such as +/- 10% or so.
Interesting. Very interesting, as I have Apple and Android devices, and my
WISP has dealt with them, and almost *every* home that complains about ****
poor WiFi reception is an Apple home where I help out.
Note that I
consider "range" to be somewhat equivalent to your "sensitivity" where
"sensitivity" is limited to receive only and does not involve the
antenna or environmental situations.
Fair enough.
Also note that anecdotal
evidence of a problem is not definitive as measurements such as
"range" and "sensitivity" tend to follow a bell curve.
I have had the two iPads tested at the Apple Genius bar, and they passed
"that" test, even though they both fail miserably at having the same WiFi
reception as all four of my Android devices have.
I'll resist the temptation to offer my opinion of Apple engineering
and RF design. Well, maybe not totally. This is my play on the
iPhone 4 antenna grip problem in 2010:
http://802.11junk.com/jeffl/cellular/cell-test.htm
Steve Jobs was right that all phones have the antenna grip problem. He
just didn't mention that the iPhone 4 had it 10 times worse than the
others.
I *knew* you had tested Apple antennas in the past!
"The worst phone I tested dropped the rx signal 16 times (-12dB). The
iPhone 4 rx signal dropped 100 times (-20dB) to 288 times (-24.6dB) .
That's a 6 to 18 times worse signal drop for the iPhone 4. Little wonder it
drops calls. My guess(tm) is that something more than detuning the antenna
is happening. I suspect that the receiver front end might be slightly
regenerative, where touching the antenna kills the regeneration and the
associated sensitivity."
How far worse? How did you measure "reception"? What were you
measuring? Using wi-fi receive signal strength from an app or
counting "bars" isn't worth much. These vary substantially between
devices and is affected by temperature.
In all cases in "my" house, I had my Android devices on a flat surface
within a foot of the iOS devices when the Android devices would easily
connect home broadband router at the far fringes of the home, while the iOS
devices failed to make any connection.
I had to solve the problem by setting up a spare WRT54G router as a wired
extension (crawling under the house and cursing Apple the entire time), so,
the fact that the WiFi reception of the iOS devices sucks compared to that
of the Android devices caused me considerable effort.
In addition, as you know, I assist my small WISP in setting up customers
and troubleshooting when they have WiFi problems. Almost invariably, the
Apple-based customers are highly non technical, so, they call up with
problems that aren't really the WISP's prerogative, such as the fact they
can't connect to either his or their routers (he insists everyone have a
router so he gives them one if they don't have their own).
Almost always, if not always, I put their iDevices next to my Android
device to test WiFi connectivity and signal strength at the distance that
the customer complains.
Even though the tools available to sniff WiFi issues on iOS devices are
downright primitive, you "can" easily see that the Android devices
"connect" to the router at distance while the iOS devices are oblivious of
the router at the same distances.
Working in the other direction, I start with the iOS device in one hand,
and the Android device in the other hand, using the primitive iOS tools and
the more sophisticated (aka modern) Android sniffing tools, so that I can
see the BSSID and perceived signal strength (in case there are multiple
SSIDs of the same name, as I have in my own setup), and almost always, if
not always, the iOS devices *drop* the connection well before the Android
devices do (I generally walk outside until the Android device drops the
connection).
After having done this so many times, whenever we get a call, we ask if
they're using Apple equipment, and, if they are, we know what's going on.
Yes of course. Since I don't like Apple, every Apple is by definition
far worse than Android. Or course, for a nominal bribe, I can reverse
the situation.
While my grandkids play games almost exclusively on the iDevices, the main
problem "I" have with Apple is that it's primitive in terms of being able
to do useful things, e.g., WiFi reception sniffers on iOS are primitive in
comparison to what's available on Android.
Jeff lives on a fault. Being honest improves my karma, and prevents
earthquakes from ruining my day.
Given the century-long cycles, you already had yours in 1989. Let someone
else have their faults!
In the past, I've offered you various ways of running a controlled
range (performance) test. The next time you get your hands on a test
device, try it. It's quite easy.
Looks like iPerf is available on iOS & Android & Windows & Linux!
https://iperf.fr/iperf-download.php
1. You will need a reasonably fast computah running iperf ver 2,
iperf3, or jperf. This turn the compoutah into an iperf server by
running just:
iperf -s
The computah should be connected via an ethernet cable to the users
router. Gigabit ethernet is nice for measuring maximum speeds, but
that's not what we're doing here.
I don't have any fast computers - but just basic laptops.
2. Next, you'll need a iperf client on the phone or tablet. There
are iperf clients for most OS's. Note that iperf2 and iperf3 are
quite different and not really compatible. If the version is not
specified, it's probably iperf2.
Android:
https://play.google.com/store/apps/details?id=net.he.networktools&hl=en
http://networktools.he.net/
https://play.google.com/store/apps/details?id=com.magicandroidapps.iperf&hl=en
IOS:
https://itunes.apple.com/us/app/he.net-network-tools/id858241710?mt=8
http://networktools.he.net/
https://itunes.apple.com/us/app/iperf-network-bandwidth-measurement/id951598770?mt=8
https://itunes.apple.com/us/app/iperf3-network-bandwidth-performance/id986846572?mt=8
PC, OS/X, Linux, etc:
https://iperf.fr/iperf-download.php
This is good that we an lay two mobile devices on a desk and run the same
iPerf utility to check performance.
Note that most Linux mutations ship with iperf2 and that iperf3 must
be installed. You can have both iperf and iperf3 installed at the
same time:
https://iperf.fr/iperf-download.php#more-recent
I'd just pick one. Probably iPerf2 for compatibility.
JAVA (runs on anything that groks Java and does pretty graphs):
https://www.rarst.net/software/jperf/
https://sourceforge.net/projects/iperf/files/
JPerf is iperf2 not 3. Version 3 is for higher speed wireless. Don't
mix versions.
I'd stick with iPerf 2.
Tutorials on iperf and jperf:
http://openmaniak.com/iperf.php
https://www.jamescoyle.net/how-to/574-testing-network-speed-with-iperf
https://www.jamescoyle.net/cheat-sheets/581-iperf-cheat-sheet
I recommend the HE (Hurricane Electric) versions which will test
either IPv4 and IPv6.
YouTube video of a typical test:
https://www.youtube.com/watch?v=4qdKgHBO_Gc
Interesting that Mike Pennacchi set up a linux box running iperf -s.
Then he runs Android iperf -c to get 10.0.10.80 as the linux box.
Then he runs Android iperf for 60 seconds & gets 50Mbps throughput.
With iPerf, not only Android but even the primitive iOS phones can be
turned into powerful network-troubleshooting tools!
This is great information!
Some notes I made from a talk on iperf and jperf:
http://802.11junk.com/jeffl/FLUG-talk-2015-03-28/iperf3%20talk.htm
You used iperf3, but I'd probably use iPerf 2 only because I want simple
compatibility with iOS and Android, Windows, and Linux.
3. Connect your test phone or tablet via wi-fi and just run a test to
see if it works. If you're running Jperf, you should see something
like this:
http://802.11junk.com/jeffl/FLUG-talk-2015-02-28/802.11gn%20direct.jpg
Note that the max speed is about 60 Mbits/sec.
Are the three graphs (purple, green, and blue) different access points?
Or are they different ports on the computer (1840, 1872, & 1860)?
If you insert a wireless repeater in between the wireless router and
the client, you get this mess:
http://802.11junk.com/jeffl/FLUG-talk-2015-02-28/802.11gn%20through%20Netgear%20repeater.jpg
Note the drastic drop in maximum speed. I'll save my rant against
mesh networks for another day.
Thanks to you, I set up a *wired* extender out of my spare WRT54G router,
as you had explained, long (long) ago, that the extender is faster, in
effect, than the repeater - even though the repeater is easier to set up
(but not easily set up on a WRT54G v5 due to lack of memory).
4. Now comes the big trick. Temporarily change the speed of your
wireless router from "automatic" to a fixed speed and/or protocol.
I've never done this. I'll have to check how to set the speed on the
Netgear WNDR2400 router and the Linksys WRT54G router. I'm sure the speed
is currently set at defaults.
For
802.11g, that would be 54 Mbits/sec. For faster protocols, it can be
faster. If you have an 802.11ac wireless router, leave both 2.4 and
5GHz on. However, if you're testing with a lesser protocol, enable
only one frequency band at a time, so that you know which one you're
testing. I would initially do the test using 802.11g and 54Mbit/sec
because higher speeds and protocols allow for fallback, which will
produce odd results.
Hmmmm....
By fixing the speed and protocol, you're eliminating the ability of
the wireless router to slow down the wireless connection speed and
thus improve the range.
Ah. That makes sense!
As you walk away from the wireless router,
instead of a general slowdown, you'll see an abrupt drop in speed,
possibly followed by a disconnect.
Eureka! That's the test I need to run!
The typical 2.4GHz 802.11g system
will go about 10 meters before the speed drops abruptly. Measure and
record this distance along with the test conditions (devices,
frequency, protocol, fixed speed, etc).
That's a PERFECT test!
My hypothesis is that the iOS devices will drop in half the distance that
the Android devices will drop - but that remains to be seen in the test.
You'll find indoor testing to vary substantially, mostly depending on
reflections and wireless router antenna positions. Outdoors works
better, but only if you don't have any interference. Try to pick an
empty channel (good luck with that).
We're in the boonies. Empty channels aren't hard to find.
Especially in the 5MHz range.
5. If you're lazy and don't want to deal with servers and iperf, you
do something similar with just ping. You still have to set a fixed
speed and protocol, but you don't get the pretty graphs and data. Just
continuously ping the wireless router. At some point, the latency
will drastically increase, followed by 100% packet loss, and possibly
a disconnect. This is not as precise as iperf because you're not
saturating the pipe with traffic, but probably good enough.
In general, for utilities, the limitation is that iTools don't allow
powerful tools, but it seems, on iOS, apparently there is a simple ping app
from MochaSoft:
https://itunes.apple.com/us/app/netw...289967115?mt=8
Android, as would be expected for a mature mobile operating system, has a
bunch of usable ping utilities, e.g., StreamSoft makes:
https://play.google.com/store/apps/d...ingtools&hl=en
6. That's all there is. The "range" of a device, which is a
measurement of the overall radio design, antenna, internal noise,
packaging, orientation sensitivity, etc quality, should give you a
clue as to relative quality of the various test devices. If
everything you test craps out at approximately the same range (using
the same speeds and protocols), then as far as I'm concerned, they're
all the same.
I would agree with you.
If the two devices under test crap out at the same time, then they're
equivalent. If the Android device craps out at twice the distance of the
iOS devices, then the hypothesis is supported.
Of course, I'd have to test multiple random devices to be sure of the data.
However, if you see substantial variations, then you
can legitimately claim that Apple and Android devices are different.
I've already many times over seen substantial variations.
So, what remains is only a more rigid test, as you have proposed.
7. Incidentally, you can also try it pointing iperf to a public
server instead of your own iperf server. Note that you'll be
measuring the speed of your internet connection, not the speed of the
wireless. I wouldn't do that for the range test.
Iperf public servers:
https://iperf.fr/iperf-servers.php
Also, if you want to be sick, try running iperf over a cellular data
connection.
Cellular is not in the cards; just WiFi.
Just do it. I didn't spend an hour writing all this so that you lean
back in your chair and deliver your "impressions" or "feelings". Such
things as range can and should be tested. If you need help, you know
where you can try to pry me out of my hole.
I will set up the iperf 2 on the various devices and test this out!